The rapid rollout of antiretroviral therapy (ART) in resource limited settings (RLS) with high rates of TB, diverse HIV subtypes and limited ART options urgently calls for research to define optimal treatment strategies. There is unprecedented interest among young health care professionals to pursue research careers in this area. The complexities and risks of conducting research in RLS require experienced investigators with protected time to adequately mentor these individuals. In this K24 renewal, Dr. Havlir proposes to train young investigators seeking academic careers in HIV related research in RLS. Dr. Havlir's initial K24 permitted her to acquire skills to design and conduct HIV research in RLS, to establish an NIH funded multidisciplinary research program in Kampala, Uganda, and to successfully compete for a T32 grant to support training of patient based research in HIV. With these resources in place, she proposes to mentor new US and international investigators through NIH funded network studies, and through "add on" studies that leverage funded cohort studies. She will also develop web based mentoring that links expertise of experienced investigators to young investigators from RLS. Dr. Havlir's research program over the next 5 years will include a leadership role in three large trials. 1. Randomized clinical trial (RCT) to define optimal timing of ART in the setting of active TB (ACTG 5221) 2. RCT to determine if early ART (CD4 >350) reduces AIDS, TB and death compared to standard of care (HPTN 052/ACTG 5245) 3. RCT to define optimal second line antiretroviral treatments. A series of pilot studies led by junior investigators to characterize drug resistance among treatment failures on first line therapy will precede the development of the study of second line therapies. The overarching goal of this proposal is to train the next generation of researchers that will address issues in RLS while addressing contemporary critical research questions. Secondarily, optimal and innovative ways to mentor young investigators working in RLS will be developed and evaluated.